Fig. 5: Genomic evolution preceding recurrent gene fusion events and the model for convergent evolution of AdhE and BdhE.

a A heatmap showing the counts/ratios of the pairs of ALDH and ADH genes adjacently coded in prokaryotic genomes, i.e., coded in the same strand of the same contigs intervened by <10 genes. The annotated numbers indicate the gene pair counts in 26,778 prokaryotic genomes, and the color scale represents the ratio of the number of adjacent gene pairs per total number of pairs co-existing in the same genomes. The ALDH and ADH orthologs in this panel correspond to ALDH/ADH orthologs included in the phylogeny of Fig. 1c and d. The number of gene pairs of AdhE- and BdhE-sister orthologs are underlined. b Multi-species comparison of gene synteny around AdhE- and BdhE-sister orthologs. Here we sampled three species in different genera for two phyla, whose genomes have AdhE- and BdhE-sister orthologs as neighbors (genes with asterisks). The colored genes are assigned KEGG ortholog identifiers. The visualization was conducted by Annoview version 2.0⁸⁹. c ColabFold prediction of heteromeric interaction between AdhE-sister or BdhE-sister ALDH and ADH proteins adjacently coded in genomes. Structural predictions were conducted by ColabFold v1.5.2. Here, we show the results of an example protein pair for the AdhE and BdhE sisters, respectively. The left heatmap and the right structures indicate the positioned aligned errors (PAEs) and the predicted structures where ALDH and ADH proteins are colored in red and cyan, respectively. d ColabFold structure prediction of heteromeric interaction between split AdhE/BdhE domains. Here, we show the results for AdhE and BdhE used in our enzymatic and structural analysis. e A schematic model of repeated genomic and structural evolution leading to extant AdhE and BdhE.